While the applications of AI/ML technology are seemingly endless, these technological offerings are often criticized for its propensity toward bias. Confirmation, survivor, and search bias are all tangible examples of how data can unintentionally be used to shape a narrative that is not an accurate representation of reality. Recent developments, however, indicate that a future of radical and positive change is on the horizon.
IBM has recently begun to utilize AI/ML to address concerns of rectifying racial inequity from a technological front. Specifically, their “Call for Code for Racial Justice” initiative is exclusively dedicated to championing advancements in tech for social good (IBM). First started amidst the prime of #BlackLivesMatter movement of 2020, this initiative continues to be the culmination of seven ongoing projects that identify and pursue areas of focus with the greatest aptitude to fight systematic, racial bias (IBM).
What does a campaign of this sort even look like in the real-world context? Where do we even start?
Access to information. The ‘power of data’ cannot fully be realized until it captures the breadth of its complexity. In other words, a large amount of data cannot be made actionable if not properly comprehended. For example, the more data that accumulates in a silo, the more untenable it will become. There can be extreme repercussions in the context of the policy and government sectors, where information cannot be used to support the interests of underserved communities (IBM). Once AI/ML has successfully streamlined the ‘readability’ of complex data sets, it will likely blossom into meaningful, impactful change for individuals from any given community.
Identifying Racial Bias. While its presence in data is not always malicious, “racial bias” can be found in a diverse range of locations. Whether accounts of crimes that took place or even police reports, there are many examples within reach where technology can at times fall short of its potential for tangible good (IBM). There are many accounts that attest to the unreliability of facial recognition algorithms, for example, due to their demonstrated lack of accuracy to identify people of color successfully (MIT). Since AI/ML models learn from a data training set, their findings will be as racist or inconclusive as its programmer. With a long history of systemic racism in the United States, we can understand the gravity of how corrupted data can be pervasive throughout many generations, if left unaddressed.
Beyond the scope. While this initiative at IBM is still in its development phase, there are many other organizations that have begun to be a force for positive, social change.
MIT’s Stephen A. Schwarzman College of Computing has already taken an invested interest in racial bias found within healthcare algorithms. Their Initiative on Combatting Systemic Racism (ICSR) is dedicated to the development of computation tools with the goal to, “help effect structural and normative change toward racial equity” (MIT). MLK visiting Professor S. Craig Watkins who helps oversee this project to ensure it adheres to a what he considers, “Ethical AI” (MIT).
MIT is not the only institution interested in bolstering racial equity in the technological space. TakeTwo and the Open Sentencing Project are other examples of programs that are dedicated in how intelligence-driven programs can be used to make an impact.
Looking ahead: the long-term impact of Ethical AI. While we are on the precipice of racial equity in the technological space, there is a lot of work that still needs to be done. Above all else, “bridging the gap” requires a community of individuals who are not only passionate about social and political impact but are also cognizant of how these decisions impact the lives of future generations.
As with anything, the first step to proper representation is visibility. It is imperative that AI/ML systems champion an inclusive, trustworthy mission if they have any potential of being a force for positive good. “Diversity” is a buzzword often thrown around without a full grasp of the power it holds when in motion. Diversity of data is the key differentiator in a model with longevity, and one that emulsifies a long history of racism and bias.
It’s often said that “a chain is no stronger than its weakest link.” Perhaps, AI/ML systems are the “link” in this context to building a brighter and more inclusive future for all.
For more information surrounding Systech Solutions’ Data Related offerings and innovations, explore https://systechusa.com/services
Modernizing Insurance business requires modernizing applications to deliver a consistent and adaptive user experience while scaling for growth and resilience.
Looking to transform your business to the cloud? Here are some questions to ask your Cloud BI provider before your move.
Are you in the pursuit of efficiency? Find out more about Managed Services, and how you can achieve a transformational outcome with your Managed Services Provider (MSP) and partner.